Discriminative Adaptation Regularization Framework-Based Transfer Learning for Ship Classification in SAR Images

Xu, Yongjie; Lang, Haitao; Niu, Lihui; Ge, Chenguang

doi:10.1109/lgrs.2019.2907139

Cited by 25 publications

(14 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Rostami et al [36] trained a DNN for SAR targets by deep transferring the weights to the target task, successfully eliminating the need for sufficient samples. Xu et al [37] employed framework-oriented transfer learning method with discriminative adaptation regularization for ship classification. In short, transfer learning can contribute to boosting performance in the case of a lack of training samples.…”

Section: Related Workmentioning

confidence: 99%

MFFA-SARNET: Deep Transferred Multi-Level Feature Fusion Attention Network with Dual Optimized Loss for Small-Sample SAR ATR

et al. 2020

View full text Add to dashboard Cite

Synthetic Aperture Radar (SAR) Automatic Target Recognition (ATR), most algorithms of which have employed and relied on sufficient training samples to receive a strong discriminative classification model, has remained a challenging task in recent years, among which the challenge of SAR data acquisition and further insight into the intuitive features of SAR images are the main concerns. In this paper, a deep transferred multi-level feature fusion attention network with dual optimized loss, called a multi-level feature attention Synthetic Aperture Radar network (MFFA-SARNET), is proposed to settle the problem of small samples in SAR ATR tasks. Firstly, a multi-level feature attention (MFFA) network is established to learn more discriminative features from SAR images with a fusion method, followed by alleviating the impact of background features on images with the following attention module that focuses more on the target features. Secondly, a novel dual optimized loss is incorporated to further optimize the classification network, which enhances the robust and discriminative learning power of features. Thirdly, transfer learning is utilized to validate the variances and small-sample classification tasks. Extensive experiments conducted on a public database with three different configurations consistently demonstrate the effectiveness of our proposed network, and the significant improvements yielded to surpass those of the state-of-the-art methods under small-sample conditions.

show abstract

Section: Related Workmentioning

confidence: 99%

MFFA-SARNET: Deep Transferred Multi-Level Feature Fusion Attention Network with Dual Optimized Loss for Small-Sample SAR ATR

et al. 2020

View full text Add to dashboard Cite

show abstract

“…Zhong et al [38] presented a simple and feasible approach by using transfer learning and achieved a good performance. Xu et al [39] proposed a differentiated adaptive regularized transfer learning framework for SAR ship classification to overcome the limitation under insufficient labeled training samples. Al Mufti et al [40] employed a pretrained AlexNet to train a multiclass SVM classifier.…”

Section: Transfer Learning Transfer Learningmentioning

confidence: 99%

Robust SAR Automatic Target Recognition Based on Transferred MS-CNN with L²-Regularization

Zhai

Deng

et al. 2019

Computational Intelligence and Neuroscience

View full text Add to dashboard Cite

Though Synthetic Aperture Radar (SAR) Automatic Target Recognition (ATR) via Convolutional Neural Networks (CNNs) has made huge progress toward deep learning, some key issues still remain unsolved due to the lack of sufficient samples and robust model. In this paper, we proposed an efficient transferred Max-Slice CNN (MS-CNN) with L2-Regularization for SAR ATR, which could enrich the features and recognize the targets with superior performance. Firstly, the data amplification method is presented to reduce the computational time and enrich the raw features of SAR targets. Secondly, the proposed MS-CNN framework with L2-Regularization is trained to extract robust features, in which the L2-Regularization is incorporated to avoid the overfitting phenomenon and further optimizing our proposed model. Thirdly, transfer learning is introduced to enhance the feature representation and discrimination, which could boost the performance and robustness of the proposed model on small samples. Finally, various activation functions and dropout strategies are evaluated for further improving recognition performance. Extensive experiments demonstrated that our proposed method could not only outperform other state-of-the-art methods on the public and extended MSTAR dataset but also obtain good performance on the random small datasets.

show abstract

“…Lang et al [14] used automatic identification system (AIS) data as the source domain to extract naive geometric features (NGFs) of the ships, and designed a multi-class adaptive support vector machine as classifier to realize knowledge transfer between two domains. Xu et al [15] proposed the method of discriminative adaptation regularization frameworkbased transfer learning (D-ARTL) which is an improvement to the original ARTL by adding a novel source discriminative information preservation regularization term to achieve a transfer from AIS domain to SAR domain. Xu et al [16] proposed the method of geometric transfer metric learning (GTML), which improves the ship classification performance in SAR domain through joint application of transfer learning and metric learning.…”

Section: Introductionmentioning

confidence: 99%

“…Lang et al [22] proposed a multi-source heterogeneous transfer learning method for SAR ship classification. Analyzing the above methods, [14], [17], [19], [21] and [22] require the support of a small number of labeled samples in SAR domain, while [15] can work without labeled samples in SAR domain, and [16] can handle both labeled samples available or unavailable situations. In terms of source domain usage, AIS data is used as the source domain by [14]- [16] and [21].…”

Section: Introductionmentioning

confidence: 99%

“…In terms of TL solution, [14] is a parameter transfer based method and the others are the methods of domain adaptation (DA), which aims at learning a well-performing model from a source data distribution on a different target data distribution. To alleviate the domain shift of data distributions across the source and target domains, [17] and [19] align the class-conditional distribution between the two domains, [15], [16] and [21] align both the marginal distribution and classconditional distribution.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Improving Deep Subdomain Adaptation by Dual-Branch Network Embedding Attention Module for SAR Ship Classification

Zhao

Lang

2022

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

Self Cite

View full text Add to dashboard Cite

This study aims at improving fine-grained ship classification performance under the condition that there is no labeled samples available in SAR domain (target domain) by transferring the knowledge from optical remote sensing (ORS) domain (source domain) which has rich labeled samples. The proposed method improves the original deep subdomain adaptation network (DSAN) by designing a dual-branch network (DBN) embedding attention module to extract more discriminative deep transferable features, thereby improving the performance of the subdomain adaptation. Specifically, we utilized a deep base network (ResNet-50) and a shallow base network (ResNet-18) to build the DBN, and embedded the convolutional block attention module (CBAM) after the first and the last convolutional layer of each branch. Extensive experiments demonstrate that the proposed method, which is termed as DSAN++, is feasible and achieves remarkable improvement than the state-of-the-art methods on the task of fine-grained ship classification.

show abstract

Discriminative Adaptation Regularization Framework-Based Transfer Learning for Ship Classification in SAR Images

Cited by 25 publications

References 15 publications

MFFA-SARNET: Deep Transferred Multi-Level Feature Fusion Attention Network with Dual Optimized Loss for Small-Sample SAR ATR

MFFA-SARNET: Deep Transferred Multi-Level Feature Fusion Attention Network with Dual Optimized Loss for Small-Sample SAR ATR

Robust SAR Automatic Target Recognition Based on Transferred MS-CNN with L²-Regularization

Improving Deep Subdomain Adaptation by Dual-Branch Network Embedding Attention Module for SAR Ship Classification

Contact Info

Product

Resources

About

Discriminative Adaptation Regularization Framework-Based Transfer Learning for Ship Classification in SAR Images

Cited by 25 publications

References 15 publications

MFFA-SARNET: Deep Transferred Multi-Level Feature Fusion Attention Network with Dual Optimized Loss for Small-Sample SAR ATR

MFFA-SARNET: Deep Transferred Multi-Level Feature Fusion Attention Network with Dual Optimized Loss for Small-Sample SAR ATR

Robust SAR Automatic Target Recognition Based on Transferred MS-CNN with L2-Regularization

Improving Deep Subdomain Adaptation by Dual-Branch Network Embedding Attention Module for SAR Ship Classification

Contact Info

Product

Resources

About

Robust SAR Automatic Target Recognition Based on Transferred MS-CNN with L²-Regularization